---
title: DTD Attribute Value Constraints
---

# DTD Attribute Value Constraints

The following are [DTD Annotations](https://unicode.org/reports/tr35/tr35.html#DTD_Annotations) that provide constraints on attribute values.
They are used internally in managing and testing the data in XML files.
Because they are for internal use for CLDR tooling, 
they are described here instead of in LDML section: [DTD Annotations](https://unicode.org/reports/tr35/tr35.html#DTD_Annotations).

They are used in DTD Annotation lines such as the following;

`<!--@MATCH:{attribute value constraint}-->`

The following describes the options available; the code implementing this is [MatchValue.java](https://github.com/unicode-org/cldr/blob/main/tools/cldr-code/src/main/java/org/unicode/cldr/util/MatchValue.java).

| Constraint                | Matches | Example | Example Match |
| ------------------------- | -------- | -------- |:--------:|
| `any`                       | any string value; often used before applying tighter constraints | -| 1438y9fbquio |
| `any/TODO`                  | placeholder for future constraints | -| uinpq43re |
| `bcp47/{subtype}`              | see bcp47 subtypes below | bcp47/cu | usd |
| `literal/{literal values}` | comma separated literals| literal/-12, 0, 87| 87 |
| `regex/{pattern}` | valid regex expression | regex/\[A-Z]{2} | BE |
| `metazone`                  | valid metazone | |Africa/Abidjan |
| `range/{start~end}` | number between (inclusive) start and end |range/0~100 | 3.2 |
| `time/{pattern}` | time or date or date-time pattern | time | y-MM-dd |
| `unicodeset/{pattern}` | valid unicodeset | unicodeset/\p{Letter} | A |
| `validity/{subtype}`       | see validity subtypes below | validity/locale | en_US |
| `version`                   | 1 to 4 digit field version | - | 35.3.9 |
| `set/{match}`              | any element of a set of elements that match \{match} | set/bcp47/tz| adalv aedxb |
| `or/{match1}\|\|{match2}`  | matches at least one of \{match1}, etc | or/bcp47/anykey\|\|literal/t | t |

## BCP 47 subtypes
These subtypes test identifiers according to the [bcp47 files](https://github.com/unicode-org/cldr/tree/main/common/validity), where key is a ukey or tkey, and type is a uvalue or tvalue.
| Constraint                | Matches | Example | Example Match |
| ------------------------- | -------- | -------- |:--------:|
| `bcp47/anykey`              | any bcp47 key | -| nu (number system)|
| `bcp47/anyvalue`            | any bcp47 value | - | roman |
| `bcp47/{key}`      | any value for that key | bcp47/nu | roman |

## Validity subtypes
Most validity subtypes are implemented in ValidityMatchValue, which test identifiers according to the [validity files](https://github.com/unicode-org/cldr/tree/main/common/validity). Each subtype may have an _idStatusList_, such as _currency_:

| Validity subtype structure                | Description |
| ------------------------- | -------- |
| `validity/currency`       | currency codes with default idStatus values |
| `validity/currency/{idStatusList}`       | specific list of idStatus values |

The optional _idStatusList_ is a list of one or more idStatus values, such as `validity/currency/regular deprecated`.
The _idStatusList_ consisting of `all` matches all idStatus values, so `validity/region/all` matches SU (Soviet Union).
The default if there is no _idStatusList_ depends on the subtype:

| Subtype                | Default idStatusList |
| ------------------------- | -------- |
| `language`, `script`    | {regular, unknown, deprecated} |
| `region`       | {regular, unknown, macroregion, special} |
| `subdivision`, `_variant_`    | {regular, unknown, deprecated} |
| `unit`, `currency`       | {regular, unknown} |

These _idStatus_ values match the values of the corresponding validity file, such as [validity/currency.xml](https://github.com/unicode-org/cldr/blob/main/common/validity/currency.xml).

* The default is all of the idStatus values that are valid for that validity file, except for deprecated. So validity/region doesn't match SU (Soviet Union).
* There is a special idStatus value `all` that includes deprecated, so `validity/region/all` matches SU (Soviet Union).
* Note that `validity/unit` tests for **long** unit values, not **short** ones.

The special matchers are:

| Subtype                | Tests |
| ------------------------- | -------- |
| `bcp47-wellformed`    | well-formed bcp47, but not for validity. |
| `locale`       | locale validity using the locale, script, region, and variant values. The idStatus is applied to each of those fields (after removing ones invalid for that field). |
| `locale-for-names`    | =locale, but also allows certain deprecated locales, and is used in matching locales names. |
| `locale-for-likely`      | =locale, also allows certain deprecated locales, and is used in matching likely subtags values. |